Validation of Esaso Classification of Diabetic Maculopathy

Abstract

Purpose To test reliability and reproducibility of ESASO morphologic OCT-based classification of diabetic maculopathy (DM). Methods This is a multi-center cross-sectional study including a coordination center (CC) and 18 participating centers (PCs). After instruction on the correct use of ESASO Classification, the validation process was carried out in two consecutive stages. In the first retrospective phase, we evaluated the concordance between PCs and CC in the staging of OCT images collected during PCs’ daily activity (608 images). In a second prospective phase, we analyzed the inter-observer agreement of staging assigned by each PCs to OCT images selected by the CC (22 images). Results The overall concordance achieved in the retrospective phase was 89.8% (Kappa = 0.83 (95% CI: 0.78–0.87); p<0.0001). In 99.5% of cases, concordance did not differ by more than one stage. In the prospective phase, PCs reached an inter-operator agreement of 93.0% (Krippendorff's Alpha = 0.953, 95% CI: 0.929–0.977, p<0.0001). Any discrepancy among the 22 images was within one stage. Conclusion The results achieved in this study confirm that ESASO OCT-based Classification can be considered as an easy and reproducible method to stage DM during clinical practice. A diffused use of a common and validated method to describe the progression of retinal damage in DM may offer several clinical and scientific advantages.